do_clean_multiv: Cleaning multivariate functional outliers

In adamethods: Archetypoid Algorithms and Anomaly Detection

Description

Cleaning of the most remarkable multivariate functional outliers. This improves the performance of the archetypoid algorithm since it is not affected by spurious points.

Usage

do_clean_multiv(data, num_pts, range = 1.5, out_perc = 80, nbasis, nvars)

Arguments

data	Data frame with (temporal) points in the rows and observations in the columns.
num_pts	Number of temporal points.
range	Same parameter as in function boxplot. A value of 1.5 is enough to detect amplitude and shift outliers, while a value of 3 is needed to detect isolated outliers.
out_perc	Minimum number of temporal points (in percentage) to consider the observation as an outlier. Needed when range=1.5.
nbasis	Number of basis.
nvars	Number of variables.

Value

List with the outliers for each variable.

Author(s)

Irene Epifanio

See Also

boxplot

Examples

## Not run: 
library(fda)
?growth
str(growth)
hgtm <- growth$hgtm
hgtf <- growth$hgtf[,1:39]

# Create array:
nvars <- 2
data.array <- array(0, dim = c(dim(hgtm), nvars))
data.array[,,1] <- as.matrix(hgtm)
data.array[,,2] <- as.matrix(hgtf)
rownames(data.array) <- 1:nrow(hgtm)
colnames(data.array) <- colnames(hgtm)
str(data.array)

# Create basis:
nbasis <- 10
basis_fd <- create.bspline.basis(c(1,nrow(hgtm)), nbasis)
PM <- eval.penalty(basis_fd)
# Make fd object:
temp_points <- 1:nrow(hgtm)
temp_fd <- Data2fd(argvals = temp_points, y = data.array, basisobj = basis_fd)

X <- array(0, dim = c(dim(t(temp_fd$coefs[,,1])), nvars))
X[,,1] <- t(temp_fd$coef[,,1]) 
X[,,2] <- t(temp_fd$coef[,,2])

# Standardize the variables:
Xs <- X
Xs[,,1] <- scale(X[,,1])
Xs[,,2] <- scale(X[,,2])

x1 <- t(Xs[,,1]) 
for (i in 2:nvars) { 
 x12 <- t(Xs[,,i]) 
 x1 <- rbind(x1, x12) 
}
data_all <- t(x1) 

num_pts <- ncol(data_all) / nvars
range <- 3 
outl <- do_clean_multiv(t(data_all), num_pts, range, out_perc, nbasis, nvars)
outl

## End(Not run)
                  

adamethods documentation built on Aug. 4, 2020, 5:08 p.m.